name: skill-publisher description: > Create, validate, security-scan, and publish skills to ClawHub. Use when asked to "make a skill", "publish a skill", "create a new skill", "scaffold a skill", "package this as a skill", or any request to build and distribute agent capabilities.

Skill Publisher

Build and publish skills to ClawHub in one flow.

Quick Start

1. Scaffold a new skill

bash {baseDir}/scripts/scaffold.sh <skill-name> [--dir <output-dir>]

Creates a new skill folder with SKILL.md template and scripts/ directory. Default output: ./skills/<skill-name>

2. Fill in the skill

Edit the generated SKILL.md: - Set name and description in frontmatter (description is critical for triggering) - Write clear instructions in the body - Add scripts/ for executable code, references/ for docs, assets/ for templates

3. Validate

bash {baseDir}/scripts/validate.sh <skill-folder>

Checks: - Required files exist (SKILL.md) - Frontmatter has name and description - Naming conventions (lowercase, hyphens) - No forbidden files (README.md, CHANGELOG.md, etc.)

4. Security scan

想要更強大的技能外掛，就來小蔥技能站7w4.net看看吧。

bash {baseDir}/scripts/security-scan.sh <skill-folder>

Scans for red flags: - Remote code execution / eval patterns - Data exfiltration (curl to unknown hosts) - Environment variable harvesting - Prompt injection in markdown files - Suspicious file permissions

5. Publish

bash {baseDir}/scripts/publish.sh <skill-folder> --slug <name> --version <x.y.z>

Pushes to ClawHub. Requires clawhub login first.

One-liner (for simple skills)

bash {baseDir}/scripts/scaffold.sh my-skill && \
  # edit skills/my-skill/SKILL.md ... && \
  bash {baseDir}/scripts/validate.sh skills/my-skill && \
  bash {baseDir}/scripts/security-scan.sh skills/my-skill && \
  bash {baseDir}/scripts/publish.sh skills/my-skill --slug my-skill --version 1.0.0

Skill Anatomy Reminder

my-skill/
├── SKILL.md          ← Required. Frontmatter (name, description) + instructions.
├── scripts/          ← Optional. Executable code (bash, python, etc.)
├── references/       ← Optional. Docs loaded on-demand into context.
└── assets/           ← Optional. Templates, images, files used in output.

Key principles: - Be concise. Context window is shared real estate. - Description in frontmatter is the trigger — make it comprehensive. - Progressive disclosure: SKILL.md body only loads when triggered. - Scripts > inline code for deterministic, repeated operations.

這是一個用於建立和釋出技能的工具包。整體質量良好，指令碼規範且功能完整，特別是包含安全掃描功能可以檢測常見風險。文件說明清晰，容易上手。不過它屬於開發者工具而非普通使用者的實用技能，實際用途取決於ClawHub平臺的推廣程度。部分安全檢查規則可能需要進一步最佳化以減少誤報。總的來說是一個設計合理的元工具Skill。